Re: poudriere: net/openldap24-server: stage/runaway , building forever

On Fri, 9 Apr 2021 12:59:59 +0200
Gary Jennejohn <gljennjohn_at_gmail.com> wrote:

> On Fri, 9 Apr 2021 12:12:45 +0200
> Guido Falsi via freebsd-current <freebsd-current_at_freebsd.org> wrote:
>
> > On 09/04/21 11:56, O. Hartmann wrote:
> > > On Fri, 9 Apr 2021 10:16:27 +0200 (CEST)
> > > Ronald Klop <ronald-lists_at_klop.ws> wrote:
> > >
> > >> Hi,
> > >>
> > >> The official pkg builders are also stuck for 14-CURRENT. Although at a
> > >> different port sysutils/msktutil.
> > >>
> > >> See main-amd64 at https://pkg-status.freebsd.org/builds?type=package
> > >>
> > >> It is stuck in "stage/runaway" for 61 hours now.
> > >> http://beefy18.nyi.freebsd.org/build.html?mastername=main-amd64-default&build=p569609_s5b3b19db73
> > >> (ipv6 only)
> > >>
> > >> NB: I'm not involved in the pkg building cluster.
> > >>
> > >> Regards,
> > >> Ronald.
> > >>
> > >>
> > >> Van: "O. Hartmann" <ohartmann_at_walstatt.org>
> > >> Datum: vrijdag, 9 april 2021 07:27
> > >> Aan: FreeBSD Ports <freebsd-ports_at_freebsd.org>
> > >> Onderwerp: Re: poudriere: net/openldap24-server: stage/runaway , building
> > >> forever
> > >>>
> > >>> On Fri, 9 Apr 2021 06:17:03 +0200
> > >>> "Hartmann, O." <ohartmann_at_walstatt.org> wrote:
> > >>>
> > >>>> Recent CURRENT host (FreeBSD 14.0-CURRENT #26 main-n245806-4d221f59b85:
> > >>>> Sat Apr  3 06:43:44 CEST 2021 amd64), poudriere CURRENT jail at
> > >>>> 14.0-CURRENT 1400007 amd64 from 2021-04-08 05:25:38. It seems that the
> > >>>> recent CURRENT does have a serious problem when building
> > >>>> net/openldap24-server. The build process gets stuck with staging and is
> > >>>> marked "runaway":
> > >>>>
> > >>>> [head-amd64-head-default] [2021-04-08_13h56m41s] [parallel_build:]
> > >>>> Queued: 1847 Built: 63 Failed: 17   Skipped: 1759 Ignored: 8
> > >>>> Tobuild: 0 Time: 13:26:35 [01]: net/openldap24-server     |
> > >>>> openldap-sasl-server-2.4.58 stage/runaway (06:28:32 / 08:41:16)
> > >>>>
> > >>>> Also, on jails (recent CURRENT) serving as OpenLDAP server (also recent
> > >>>> taken from git /usr/ports, branch main), run into a serious problem
> > >>>> starting slapd, when starting slapd and the process is reporting
> > >>>> checking configuration, it freezes forever. Putting slapd into debug
> > >>>> mode doesn't help, since the freeze is quite early.
> > >>>>
> > >>>> Does anybody know what the reason for this strange behaviour is on
> > >>>> CURRENT? All CURRENT servers are affected (almost all the same revision
> > >>>> as shown above)?
> > >>>>
> > >>>> Thanks in advance,
> > >>>>
> > >>>> O. Hartmann
> > >>>
> > >>> Short update, another host is stuck at the very same point, host's
> > >>> CURRENT is at FreeBSD 14.0-CURRENT #2 main-n245870-86a52e262a6: Wed Apr
> > >>>  7 13:57:20 CEST 2021 amd64, it's jails is taken from the same source.
> > >>>
> > >>> The process is stuck at staging and took 34 hours ... never seen before:
> > >>>
> > >>>
> > >>> [...]
> > >>> [09:05:25] [03] [02:13:44] Finished net/openldap24-server |
> > >>> openldap-sasl-server-2.4.58: Failed: stage/runaway load: 10.39  cmd: awk
> > >>> 24374 [running] 0.06r 0.00u 0.00s 0% 3420k [headamd64-head-default]
> > >>> [2021-04-07_12h26m18s] [parallel_build:] Queued: 3298 Built: 2123
> > >>> Failed: 7 Skipped: 1161 Ignored: 7    Tobuild: 0     Time: 40:52:34
> > >>> [03]: net/openldap24-server     | openldap-sasl-server-2.4.58
> > >>> stage/runaway (31:48:30 / 34:01:11) [40:52:52] Logs:
> > >>> /pool/poudriere/data/logs/bulk/headamd64-head-default/2021-04-07_12h26m18s
> > >>> _______________________________________________
> > >
> > > [...]
> > >
> > > It seems, that jails on 14-CURRENT do have a strange and malfunctional
> > > behaviour now.
> >
> > Most probably related to commit d36d68161517 check the thread at [1].
> >
> > A solution is being discussed in D29623 at [2].
> >
> >
> > [1]
> > https://lists.freebsd.org/pipermail/dev-commits-src-all/2021-April/005159.html
> >
> > [2] https://reviews.freebsd.org/D29623
> >
>
> Just FYI I was also was seeing strange hangs with dbus (polkitd was
> timing out) and firefox was hung waiting for urwlck.  I also saw
> timeouts waiting for Xorg to shutdown (startx ended up kiliing it).
>
> I applied the patch which kib posted and everything returned to normal.
>

I also applied the patch mentioned bz Guido Falsi above and services like
openLDAP and Apche start now as expected. It seems the patch works on amd64, at
least for some boxes (those I've already pacthed right now).

Thank you very much.

Kind regards,

O. Hartmann